APPLICATION DATA
AD353-129
Procidia Control Solutions
Feedforward Control
This application data sheet describes several
feedforward control methods. Any of these
configurations can be readily developed within a
Siemens 353 controller. 1
Feedforward control is a method of compensating for
process load disturbances before they can affect the
primary control variable. This control strategy is
particularly beneficial for critical process control loops
that are often slow to respond to load disturbances.
By reducing the effect of these load disturbances,
feedforward control minimizes the magnitude and
duration of the control errors that would normally
occur without this control strategy.
In the heat exchanger example shown in Figure 1,
exit temperature is the primary control variable. The
temperature controller (TC) manipulates the flow of
Steam
TC
TT
Feed Exit
Temperature
Heat
Exchanger
T potential for unstable operation. Since all feedback
Condensate
Figure 1 Heat Exchanger Temperature Control
1
See Applications Support at the back of this publication
for a list of controllers.
Rev 2
April 2012
steam to the heat exchanger to control the
temperature at setpoint. An increase in the feed flow
to the heat exchanger will cause a decrease in the
exit temperature. The temperature controller
responds to this control error by increasing the steam
flow as required to return the temperature to setpoint.
The change in feed flow represents a change in load
on the heat exchanger and is referred to as a load
disturbance. Other load disturbances that can affect
the exit temperature are changes in inlet
temperature, ambient temperature, and steam
pressure.
The temperature controller can compensate for any
and all load changes. That is the essential
advantage of feedback control. No matter what is
causing a change in exit temperature, the controller
will detect the error and change the steam flow as
required to reduce the error to zero. Figure 2 is an
example of the response of a single-loop PI controller
to a change in feed flow. As expected, the controller
returns the exit temperature to setpoint after the load
disturbance from the change in feed flow.
The disadvantage of feedback control is that it can
detect the load disturbance only after the disturbance
has occurred. Due to the dynamics of the process,
the load disturbance will go undetected for some time
before the exit temperature begins to change, and
then additional time will be required for the exit
temperature to respond to the change in steam flow
manipulated by the controller to compensate for the
new load.
Another disadvantage of feedback control is the
control loops will cycle if tuned to respond too quickly,
the speed at which the temperature controller can
compensate for the disturbance is restricted by the
stability limit of the control loop.
AD353-129

Exit Temperature

Steam Flow

Feed Flow

Figure 2 Single-Loop Response to Feed Flow Change

Figure 3 shows an example of feedforward control scaling and dynamic compensation, it is theoretically
applied to the heat exchanger temperature control possible to match the change in valve position to the
loop. The load disturbance, caused by a change in change in feed flow so that there is no effect
feed flow, is measured and added directly to the exit whatsoever on the exit temperature.
temperature controller output. This allows an
increase in feed flow to provide an immediate It should be noted that feedforward is an open loop
increase in steam flow. With appropriate signal control strategy. The change in the manipulated
variable (steam flow) has no effect on the load
Steam variable (feed flow). Therefore, it is not possible for a
feedforward loop to become unstable.
+ TC
For feedforward to be effective, it must be configured
+
so that the change in load moves the manipulated
variable in the right direction, the right amount, and at
the right time. The right direction can be deduced
from an understanding of the process. The right
TT
FT amount and the timing may be estimated from
process knowledge, but is more often determined by
empirical testing of the control loop and its response.
Feed Exit
Temperature
If feedforward adds too much steam for an increase
in feed flow, the exit temperature will increase, and
Heat
Exchanger
there will be a control error in the opposite direction
T from that caused by the load change. The control
configuration must include a gain adjustment so that
the magnitude of the change in the manipulated
Condensate variable can be adjusted relative to the change in the
load variable.
Figure 3 Heat Exchanger Steady State
Feedforward Control A similar problem occurs if the change in steam flow
is too soon or too late. Since exit temperature will
respond at different rates to a change in feed flow
and a change in steam flow, it may be necessary to
slow down (or speed up) the response of the

2

manipulated variable to a change in feed flow. The
control configuration must include dynamic
compensation to synchronize the response of the
manipulated variable with the response of the load
variable. This is accomplished by adding lead or lag,
and perhaps dead time, to the feedforward signal.
Although feedforward control provides significant
benefits, it is not practical to control a process by
feedforward alone. Feedforward can only
compensate for measured load variables. It is not
practical to measure every conceivable load variable
that can disturb a process. In addition, it is not
possible to provide perfect feedforward compensation
for the load variables that are being measured. Even
the most sophisticated feedforward control scheme is
not capable of controlling the process variable exactly
at setpoint. Therefore, it is always necessary to use
feedback control in conjunction with feedforward
control to “trim” or “bias” the feedforward model.
For feedforward control, consider one of the
techniques below, which are discussed in detail in the
remainder of this publication.
• Steady State Feedforward
• Impulse Feedforward
• Model Predictive Feedforward
STEADY STATE FEEDFORWARD
The steady state feedforward configuration is shown
in Figures 4 and 5. The PID function block in loop 1
manipulates its output to control the exit temperature
at the setpoint value from the SETPT block.
The feed flow signal is the load variable for
implementing feedforward control. In this application
example the feed flow signal is acquired using
Ethernet I/O and brought into the 353 (loop 2)
through the AIE1 function block. The lead/lag block
(LL1) provides adjustable dynamic compensation. It
can be configured to provide lag or lead
compensation. A dead time block (DTM1) has also
been included for any dead time compensation that
may be required (see AD353-127 for dead time
compensation techniques).
The proportional-only controller (PD) in loop 2
provides signal scaling and bias for the feedforward
signal. If the PD controller is configured for direct
action, the output signal is determined by the
following equation:
3
AD353-129
CO = PG( P − S ) + MR
where:
CO is controller output
PG is proportional gain
P is process variable (feed flow)
S is setpoint from SETPT block
MR is manual reset
The values for S and MR are initialized whenever the
loop is switched to the manual mode. The “not auto”
(NA) status signal from the A/M transfer switch in
loop 1 will force the SETPT block in loop 2 to track
the feed flow signal and the manual reset of the PD
block to track the output of the A/M block. When the
loop is returned to auto, the values of S and MR are
held at the values of their respective track variables
at the moment the transfer is made. Since P=S and
MR equals the valve signal, the output of the PD
block is initialized at the current valve output. Any
subsequent change in the feed flow will change the
output of the PD block by an amount equal to PG(P –
S).
Math block MTH1 in loop 1 adds the feedforward
component from the PD block to the output of the PID
exit temperature controller. Except as filtered by the
lag block (LL1), this provides an immediate change in
the valve signal for changes in feed flow. The
characterizer block (CHR1) is available to
compensate for any nonlinearity in the relationship
between valve position and steam flow to the heat
exchanger.
Notice that MTH1 is configured with a – 50% bias
value. This allows the PID controller output to trim
the feedforward component in both a positive and
negative direction. When the PID block output is at
midscale (50%), the controller output signal is
cancelled by the bias and the feedforward component
is passed through without bias. However, the
controller can adjust the valve loading signal up or
down by varying its output above or below midscale
respectively.
The A/M transfer block in loop 1 allows the operator
to switch between auto and manual modes of
operation. To provide bumpless transfer from
manual to auto, the PID controller must be initialized
at a value that will allow the MTH1 function block to
track the output of the A/M block in the manual mode.
This is accomplished by “back calculating” the
appropriate reset feedback in math block MTH2. In
manual, the PID controller tracks the feedback signal
provided by the MTH2 block and the PD controller
AD353-129

Figure 4 Exit Temperature Controller with Steady State Feedforward Control (CF353-129SS)

Figure 5 Steady State Feedforward Computation for Exit Temperature Controller (CF353-129SS)

4

tracks the output of the A/M block. This forces the
output of the PID block to midscale (50%); which
forces the output of the MTH1 block to match the
output of the A/M block.
An example of the response to a feed flow change for
this example is shown in Figure 6. Compared with
the single loop only response in Figure 2, the
feedforward compensation has reduced the effect of
a feed flow load disturbance significantly. Depending
on the dynamic interaction between the feed forward
variable (feed flow) and the controlled variable (exit
temperature), the amount of improvement may be
limited. The tuning of the feedforward compensation
is also critical to the performance of the feedforward
compensation.
PG = 0.58, TLAG = 0.01 min, TLEAD = 0.18 min.
Figure 6 Response to a Feed Flow Change with Feedforward
Tuning
In a feedforward loop, the primary feedback controller
must be tuned before tuning the feedforward
compensation. To minimize the effect of feedforward
while tuning the primary controller, set the lag block
(LL1) time constant (TLAG) and the PD controller
gain (PG) to their minimum values. The feedback
controller can then be tuned using any of the
standard methods including AUTOTUNE in a 353.
Next, the feedforward gain and time constant are
tuned for the best feedforward compensation. With
exact compensation, a change in feed flow would
have little or no effect at all on the exit temperature.
However, exact compensation is often not possible
because a simple lead or lag response cannot
compensate for the more complex process dynamics
normally encountered. Therefore, the objective is to
5
AD353-129
minimize the change in the controlled variable for a
change in the load variable.
For tuning the feedforward response, it is most
convenient if the load can be adjusted to change the
feedforward variable. When this is not possible, it is
necessary to observe and, preferably, record the
effect of changes as they occur. A record of the
valve position is particularly useful for tuning.
Figure 7 shows four feedforward tuning examples
and the resulting valve positions. In all cases, a step
change occurred at the instant the valve plot
undergoes a step change upward. The horizontal
dashed line represents the new steady state valve
position that returns the controlled variable to
Exit Temperature
Steam Flow
Feed Flow
setpoint. Dynamic compensation has been
minimized for these examples.
Example A shows the effect of too much feedforward
gain. Feedforward moves the valve too far initially,
and the feedback controller must reduce its output to
compensate for the over-correction. The only way
the controller can accomplish this is to integrate a
control error caused by the initial overcorrection.
Example B shows the effect of too little feedforward
gain. Feedforward does not move the valve far
enough initially, and the feedback controller must
increase its output to compensate for the under-
correction. Again, integrating a control error is
required.
In example C, the feedforward gain is set correctly.
The initial valve position achieved by feedforward is
the same position that satisfies the feedback
AD353-129
controller at steady state. However, the transient
increase in valve position after the initial step
indicates that the controlled variable did not remain at
setpoint initially. Feedforward was too late to prevent
the transient control error. Clearly, feedforward
cannot be expected to respond before the load
change occurs. However, lead compensation
provides an initial kick to minimize the effect of being
late.
In example D, the feedforward gain is also set
correctly. However, in this case, the transient
decrease in valve position indicates that the
feedforward response was too early. Lag
compensation can minimize the effect of being too
early.
Typically, the feedforward gain (PG) should be
adjusted to the correct value before adjusting the
lead or lag settings. Then, find the dynamic
compensation that minimizes the transient effect.
Figure 7 Feedforward Tuning Examples
6
The dynamic compensation will rarely be perfect due
to the complex nature of the process dynamics.
Comparison with Other Methods
Steady state feedforward is relatively simple to
implement and tune. However, it is important to
realize that this method of feedforward control can
prevent the feedback controller from stroking the
valve completely. If, for example, the feedforward
component is contributing more than 50% to the
valve signal, the feedback controller will not be able
to bias the feedforward sufficiently in the negative to
close an air-to-open valve. This does not normally
occur unless the feedforward component is badly
misadjusted.
If a sustained control offset develops due to the
biasing problem described above, momentarily switch
the loop to manual to re-initialize the controllers and
then switch the loop back to auto to resume control.
 AD353-129

IMPULSE FEEDFORWARD Figure 9 shows the combination of a deviation

amplifier (DAM_) block and a lag (LL_) block that is
Impulse feedforward derives its name from the fact used to implement the impulse feedforward. The
that the feedforward (load) variable generates an load variable is connected to the A and B inputs (B
impulse that provides a transient bias to the controller via the lag block), and the controller output is
output. The valve loading signal is the sum of the connected to the C input. At steady state, A and B
PID controller output and the feedforward impulse. cancel one another, and the output of the deviation
This feedforward technique is shown in Figure 8. amplifier is simply the value of the C input (controller
output).
Steam Reset
Feedback For a step change in the load variable, there is an
immediate change (impulse) in the output of the
IMP + TC deviation amplifier. However, as the lag responds to
+ the step change, the A input is increasingly cancelled
by the B input, until at steady state the load variable
contributes nothing to the output of the deviation
amplifier. The initial impulse decays to zero, and the
TT
output of the deviation amplifier returns to the value
FT
of the controller output (input C).

Exit An impulse that decays to zero does not provide the

Feed
Temperature steady state change in the manipulated variable that
is necessary to compensate for the load disturbance.
Heat However, as shown in Figure 10, the output of the
Exchanger deviation amplifier provides external reset feedback
T to the controller. This allows the impulse to drive the
reset (integral) component and force the controller
output to a new steady state operating value. If the
lag time (TLAG) is set equal to the integral time, the
Condensate reset component will complement the decaying
impulse. The sum of these two transients will provide
Figure 8 Impulse Feedforward a sustained step change in the output of the deviation

Load Controller
Variable Output 100%

Gx L

C DEV AMP
A Output
DEV TL
AMP
Load Variable
B

Lag L

0
GAIN(A-B) + C t

Figure 9 Impulse Function

7
AD353-129
Load
Variable
FB 100%
PID
Gx
C Impulse
A
DEV
AMP
B
Lag
0
G(A-B) + C
Figure 10 Impulse Plus Reset Feedback
amplifier. This achieves a steady state change in
valve loading without integrating a control error or
adding a steady state bias.
Impulse feedforward can also provide dynamic
compensation by setting the lag time (TLAG) to be
less than or greater than the integral setting (TI) of
the controller. For lead response, TLAG must be less
than TI, and for lag response, TLAG must be greater
than TI.
Configuration
Figure 11 shows the configuration for impulse
feedforward. The PID controller manipulates its
output to control the exit temperature at setpoint.
The feed flow signal is the load variable for
implementing feedforward control. This signal is
applied directly to one side of a deviation amplifier
(DAM1) and indirectly through a lag block (LL1) to the
other side of the deviation amplifier. If the application
requires an inverse relationship between the valve
signal and the load variable, these connections to the
deviation amplifier must be reversed. A change in
feed flow generates an impulse as described above.
The A/M transfer switch allows the operator to switch
between auto and manual modes of operation. To
provide bumpless transfer from manual to auto, the
auto status (AS) signal from the A/M block forces the
PID block to track the valve signal and the lag block
(LL1) to bypass the lag function.
8
DEV AMP
Reset Output
L
TL
Load Variable
L
t
If desired, a characterizer block can be inserted in the
valve signal, as shown in Figure 4 for the steady
state feedforward configuration. However, the
external feedback for the PID block must “tee off”
upstream of the characterizer block.
Tuning
Impulse feedforward is tuned in much the same way
as steady state feedforward. The main difference is
the manner in which dynamic compensation is
accomplished when using impulse feedforward.
With the gain (GAIN) of the deviation amplifier
(DAM1) and the time constant (TLAG) of the lag
block (LL1) set at minimum values, the feedback
controller (PID) can be tuned using any of the
preferred methods.
Before doing tests to determine the feedforward gain,
the lag time constant should be set equal to the
integral time constant of the controller ( TLAG = TI).
This adjusts the impulse for no lead or lag. Use
examples A and B in Figure 7 as a guide to find the
correct feedforward gain setting.
With the correct feedforward gain, use examples C
and D in Figure 7 to determine the need for lead or
lag action. If lead is required, decrease the lag time
constant (TLAG) and if lag is required, increase the
lag time constant.

Figure 11 Impulse Feedforward Configuration (CF353-129IM)
The steady state gain (G) of the impulse function is
affected by both the deviation amplifier gain
adjustment (GAIN) and the ratio of the integral and
lag time constants as follows:
TLAG
G = GAIN ( )
TI
To retain the effective feedforward gain for changes
in TLAG, recalculate the gain of the deviation
amplifier (GAIN) per the following equation:
TI
GAIN = G( ) controller “trims” the model to compensate for
TLAG
Suppose, for example, that the feedforward gain was
set for 1.2 when TLAG = TI. If the lag time (TLAG) is
readjusted to 0.8TI to achieve a leading response,
the GAIN of the deviation amplifier should be
readjusted to (1.2)/(0.8) or 1.5 to retain the original
feedforward gain. It will be necessary to readjust
GAIN for every value of TLAG that is tried while
conducting experiments to determine the optimum
value of TLAG.
Comparison with Other Methods
Impulse feedforward is accomplished with a simple
configuration. It also eliminates completely the
potential bias problem that can occur with steady
state feedforward.
9
AD353-129
The only disadvantage of impulse feedforward is the
potentially disconcerting interaction among the tuning
constants during the tuning process. The tuning
procedure is essentially the same in both cases, but
the interaction must be understood to avoid
frustration.
MODEL PREDICTIVE FEEDFORWARD
Model predictive feedforward uses a rigorous
mathematical model of the process to calculate the
value of the manipulated variable as a function of one
or more load variables. The primary feedback
modeling errors or unmeasured load disturbances.
As shown in Figure 12, both the feed flow and the
feed temperature are measured. The output of the
temperature controller (TC) represents the exit
temperature. In this application, the heat required to
raise the feed flow to the desired exit temperature
can be calculated using the following equation.
Q = c p F f Δ`T
where:
Q is heat transfer rate [e.g. BTU/hr]
cp is specific heat [e.g. BTU/(lb·ºF)]
Ff is feed flow (mass) [e.g. lb/hr]
∆T is change in temperature [e.g.·ºF]
AD353-129
The heat supplied by the steam can be calculated
using the following equation:
Q = λFS
Where:
Q is heat transfer rate [e.g. BTU/hr]
λ is the latent heat of steam [e.g. BTU/lb]
Fs is steam flow (mass) [lb/hr]
This equation is used to calculate the value of the
manipulated variable in a model predictive
feedforward configuration. For other applications, it
will be necessary to use whichever first principles
(heat balance, material balance, etc.) are appropriate
to derive a rigorous mathematical model.
Steam
Ff T T
X + TC
_
FT TT TT
Feed Exit
Temperature
Heat
Exchanger
T block to calculate the correct external feedback for
Condensate
Figure 12 Model Predictive Feedforward
Configuration
The configuration for model predictive feedforward is
shown in Figure 13. The math block (MTH1) is used
to calculate the steam flow as a function of feed flow,
feed temperature, and exit temperature setpoint.
Note that the output of the exit temperature controller
is used instead of the actual exit temperature setpoint
as a convenient method of providing feedback trim to
the feedforward calculation. Otherwise, another math
block would be required to implement feedback trim.
10
In addition, if both the setpoint and the controller
output are used in the feedforward calculation, there
will be a double correction whenever the setpoint is
changed. This configuration avoids that problem. If
the model provides perfect steady state feedforward
compensation, the temperature controller will be
satisfied when the output equals the desired exit
temperature. Otherwise the PID block will bias the
∆T calculation as required to drive the temperature to
setpoint.
The equation derived above can be implemented
using a single configurable math block (MTH_). A
lead/lag block (LL_) provides dynamic lead or lag
compensation. A dead time block (DTM_) provides
any dead time compensation that may be required.
Although all load variables may require dynamic
compensation, it is usually more critical on load
variables that move more quickly such as flow.
The A/M transfer switch allows the operator to switch
between auto and manual modes of operation. To
provide bumpless transfer from manual to auto, the
PID block must be initialized at a value that will allow
the math block (MTH1) to track the output of the A/M
block in the manual mode. This is accomplished by
“back calculating” the appropriate reset feedback in
math block MTH2. In manual, the PID block tracks
the feedback signal provide by MTH2 and the AS
status from the A/M block bypasses the dynamic
functions of the LL_ and DTM_ blocks.
If desired, a characterizer block can be inserted in the
valve signal, as shown in Figure 4 for the steady
state feedforward configuration. However, input A to
MTH2 must “tee off” upstream of the characterizer
the PID block. Alternatively, if a steam flow
measurement is available, this configuration can be
modified to provide a setpoint to a secondary steam
flow controller in cascade. See AD353-128 for more
information on Cascade Control.
The Siemens 353 controller uses real (floating point)
numbers between blocks. This enables engineering
units to be used for block interconnection signals.
The PID block output is scaled to represent the exit
temperature with a range of 0 to 200ºF. The signal
out of the MTH1 block represents steam flow. Scaler
block SCL1 provides a steam flow range to the
analog output block so it can convert the steam flow
range of 0 to 6000 lbs/hr to 0 to 100% valve signal.
The MTH2 block output signal represents exit
temperature for reset feedback to the PID block. The
calculations for the MTH_ blocks were derived as
follows:
 AD353-129

Figure 13 Model Predictive Feedforward Configuration (CF353-129MP)

CP
MTH1 Fs = ( )(60 ρF )(T
f e − Tf ) This method requires more design effort than either
λ the steady state or impulse feedforward methods.
However, the effort required will usually be rewarded
MTH2 T e = ( λ )( Fs
) + Tf by significant improvement in control loop
performance.
Cp 60 ρ F f
APPLICATIONS
Ff (Feed Flow) ................ 0 to 400 GPM
The feedforward control techniques described in this
Tf (Feed Temp) ............... 0 to 200 ºF
publication are applicable to any process control loop
Te (Exit Temp)................. 0 to 200 ºF
with measurable load disturbances. Examples are
Fs (Steam Flow) .............. 0 to 6000 lbs/hr
boiler drum level control, composition control in
Cp (specific heat) ............ 0.8 BTU/lb (feed) ºF
distillation columns, and pH control in waste water
ρ (density)........................ 8.25 lb (feed)/ gallon
treatment.
λ (latent heat of steam).... 853.5 BTU/lb of steam
Application Support
Tuning
User manuals for controllers and transmitters,
With model predictive feedforward, the mathematical
addresses of Siemens sales representatives, and
model establishes the relationship between the
more application data sheets can be found at
manipulated variable and the load variables.
www.usa.siemens.com/ia. To reach the process
Therefore, there is no feedforward gain adjustment,
controller page, click Process Instrumentation and
except that which is already built into the model. It is
then Process Controllers and Recorders. To select
necessary to determine only the type of dynamic
the type of assistance desired, click Support (in the
compensation required and to tune it in much the
right-hand column). See AD353-138 for a list of
same manner as described for steady state
Application Data sheets.
feedforward.
The configuration(s) shown in this publication were
Comparison with Other Methods
created in Siemens i|config™ Graphical
Configuration Utility. Those with CF353 in
Model predictive feedforward is the most
parenthesis in the Figure title are available using the
sophisticated of the three methods described and has
above navigation, then click Software Downloads >
the potential to provide the best performance. It can
353 Feedforward Control (Reference AD353-129).
compensate for more than one load variable and is
based on fundamental knowledge of the process.

11
AD353-129

The configuration(s) can be created and run in a:

• Model 353 Process Automation Controller
• Model 353R Rack Mount Process Automation
Controller*
• i|pac™ Internet Control System*
• Model 352Plus™ Single-Loop Digital Controller*
* Discontinued model

i|pac, i|config, Procidia, and 352Plus are trademarks of Siemens Industry, Inc. Other trademarks are the property of their respective owners.
All product designations may be trademarks or product names of Siemens Industry, Inc. or other supplier companies whose use by third
parties for their own purposes could violate the rights of the owners.

Siemens Industry, Inc. assumes no liability for errors or omissions in this document or for the application and use of information in this
document. The information herein is subject to change without notice.

Siemens Industry, Inc. is not responsible for changes to product functionality after the publication of this document. Customers are urged to
consult with a Siemens Industry, Inc. sales representative to confirm the applicability of the information in this document to the product they
purchased.

Control circuits are provided only to assist customers in developing individual applications. Before implementing any control circuit, it should
be thoroughly tested under all process conditions.

Copyright © 2012, Siemens Industry, Inc.

12